Re: [DEP9] call for testing of reconf-inetd (update-inetd replacement)

[Serafeim Zanikolas <sez@debian.org>]

Hi Ian,

First of all, thanks for the feedback.

On Mon, Feb 13, 2012 at 12:36:40PM +0000, Ian Jackson wrote:
> Serafeim Zanikolas writes ("[DEP9] call for testing of reconf-inetd (update-inetd replacement)"):
> > To test it, in a nutshell:
> > 
> >     - make your package install xinetd.conf(5) fragment files in
> >       /usr/share/reconf-inetd/, one file per service that needs an entry in
> >       inetd.conf
> > 
> >     - run reconf-inetd as root
> > 
> >     - check whether an entry has been added in /etc/inetd.conf.
> 
> Thanks.  update-inetd has needed a revamp for quite a while.  I'll
> hopefully have a chance to look at this properly at some point, but in
> the meantime:
> 
> Are you sure putting the fragments in /usr/share is the best 
> approach ?  It seems to me that it might be better to put them in
> /etc/, where the sysadmin can edit them if they want to change the
> behaviour but still get the benefit of the automatic addition/removal.

Any local sysadmin changes must be done in inetd.conf, as always.

The choice of /usr/share/ follows from two of the requirements I have set from
the beginning for DEP9 [0]:

- the standard configuration files of inetd and xinetd must remain the
  authoritative files;
- the solution must not change the way system administrators configure inetd

/usr/share/reconf-inetd fragments are not configuration files; they're just
input to reconf-inetd. You can think of them as the file-based equivalent of
command line arguments to update-inetd

cheers,
sez

[0] an older proposal of mine in -devel was rejected (rightfully so) precisely
    for not meeting these requirements